Recent Developments in the Prevention and Treatment of Missing Data

Area of Science:

• Biostatistics
• Clinical Research Methodology
• Data Science

Background:

• The National Research Council (NRC) identified a critical need for improved methods and software for handling missing data in research.
• The Drug Information Association Scientific Working Group (DIASWG) was formed to address these recommendations.

Purpose of the Study:

• To distill the NRC's 18 recommendations into a practical framework for managing missing data.
• To provide guidance on conducting sensitivity analyses for assessing the robustness of research inferences.
• To introduce freely available software tools for missing data analysis.

Main Methods:

• Distilled 18 NRC recommendations into three core pillars for addressing missing data.
• Utilized sample datasets to demonstrate the application of sensitivity analyses.
• Developed and made publicly available a suite of software tools for missing data analysis.

Main Results:

• Established a three-pillar framework: clear objectives/estimands, missing data prevention, and primary/sensitivity analyses.
• Demonstrated how sensitivity analyses enhance the assessment of inference robustness.
• Provided accessible software tools for implementing these methods.

Conclusions:

• The proposed framework and tools facilitate more rigorous handling of missing data in research.
• Sensitivity analyses are crucial for evaluating the reliability of conclusions when data are missing.
• Publicly available software promotes wider adoption of best practices for missing data management.